1. Field of the Invention
The present invention relates generally to insulated gate field effect transistors and more specifically to an improved insulated gate field effect transistor.
2. Prior Art
Insulated gate field effect transistors of the prior art are generally illustrated in FIGS. 1, 2, and 3.
The device shown in FIG. 1 is a planar IGFET consisting of diffused source and drain regions 14 formed in a body 12 of the opposite conductivity type. The conductive control electrode or gate 18 is separated from the source, drain and body by a thin insulating layer 20. FIG. 1 shows an N channel IGFET; however, the basic principles of operation are independent of polarity type.
The improvements to be described in this invention consist of means to minimize both the channel length L and the effective depth of the source and drain regions x.sub.JS and x.sub.JD. In the device shown in FIG. 1, x.sub.JS and x.sub.JD are determined by the depth of the N+ diffusion used to fabricate the source and drain regions. The source-drain spacing determines the channel length L. The minimum channel length of the device shown in FIG. 1 is determined by the photolithographic tolerances that can be held during definition of the source and drain regions.
Operating speeds are inversely proportional to channel length; therefore it is desirable to make the channel as short as possible. For the device shown in FIG. 1, the minimum channel length is, in practice, limited by the resolution obtainable in the photolithographic process.
For very short channel devices, a second phenomenon may determine the minimum channel length. For a fixed drain bias, there will be a space charge region associated with the drain-substrate junction. If the width of this space charge region is greater than L, the gate electrode may not be able to effectively control the conductivity of the channel region. Although an exact analysis of this "punch-through" condition is quite involved, in general the effect is minimized by making x.sub.JS and x.sub.JD as small as possible. In the realization of the device shown in FIG. 1, x.sub.JS and x.sub.JD are minimized by ion implantation and by the use of slow diffusing impurities such as arsenic or antimony.
Improvements in the prior art of FIG. 1 are shown in FIG. 2 and FIG. 3. In both devices, the channel length is defined by controlling the depth of a conducting layer diffused from the surface. This gives shorter practical channel lengths than are usually obtained with the device shown in FIG. 1.
The FIG. 2 device includes a source layer 10, a body region 12, and a surface drain region 14. A gate material 18 is formed in groove 16 and separated therefrom by an insulated layer 20. The channel of the VMOS device has a length L between the surface drain 14 and the source layer 10. The depth x.sub.JD of the drain 14 is defined from the channel. The prior art device of FIG. 2 has a length L and drain junction depth x.sub.JD.
An improvement over the prior art device of FIG. 2 is illustrated in FIG. 3. The prior art device of FIG. 3 includes a region 22 extending into layer 12 of the same conductivity type as the drain region 14. Region 22 reduces the length of the channel but increases the depth of the drain region relative to the gate surface. Similarly, it should be noted that FIG. 3 illustrates a generally U-shaped vertical groove compared to the V-shaped vertical groove of FIG. 2. This specific shape of the groove is interchangeable and would depend on the method of fabrication. The impurity regions of the prior art devices are formed before the groove and gate. The subsequent processing increases the depth of the regions and consequently alters the channel length.
For the devices shown in FIGS. 2 and 3, the minimum channel length is limited by punchthrough. The device shown in FIG. 3 minimizes the punchthrough effect somewhat by allowing the drain-substrate space charge region to spread partially through the lightly doped region 22. As shown in the figures, the effective depth of the drain junction is determined by the distance from the edge of the groove to the edge of the N+ region, and in practice, is limited by a combination of photolithograph, alignment and etch tolerances.
Prior art devices have thus failed to simultaneously minimize channel length L and drain depth x.sub.JD to produce a high performance IGFET.